In order to bring a method that provides absolute quantification, like digital PCR, to the clinic, it is important to develop appropriate quality control methods to demonstrate that results are not only precise, but also accurate. The dPCR method can provide a readout of the exact number of nucleic acids in a reaction droplet, but it has also been difficult to pinpoint errors caused by variability in droplet size, non-uniformity of block heating in a thermal cycler, or other parameters.
The National Measurement Institute of Australia has now created a reference material for digital PCR that can be used to check the behavior of each well in a 96-well dPCR format. As described in Analytical Chemistry last month, the material was ultimately used in an inter-lab comparison in seven labs and was tested on the Bio-Rad Droplet Digital PCR instrument, but could also be more broadly applicable.
Bio-Rad has worked with researchers at NMI previously but did not participate in this study, although George Karlin-Neumann, the director of Bio-Rad's digital biology center, described it in an email as "well-conceived and executed."
The study pointed out that "digital PCR can provide unprecedented levels of precision, accuracy, and resolution for quantification of nucleic acids," and that this, together with development and availability of affordable instrumentation, have been the main reasons for the rapid expansion of the technique.
Study author Leonardo Pinheiro, a senior scientist at NMI's bioanalysis group, said the original motivation of his team was simply to develop a reference material that could be used in a well-based format. However, the study also happened to uncover an interesting edge effect in a thermal cycling instrument used by one of the labs.
By chance, the particular target assay the NMI group initially chose happened to be sensitive to inaccurate denaturation cycling temperatures, Pinheiro said. Specifically, the assay used a challenging template with a high GC content. In one instrument tested with the assay, the researchers noted variability that led them to hypothesize the thermal cycling block was not heating the edges as much as the middle, thus causing readings that were not actually due to variation in absolute amount of DNA in each well.
A high-precision acoustic dispensing method was used to generate the reference material with a "certified reference value expressed in amount of DNA molecules per well," according to the study. This method was chosen over typical pipetting-based techniques because it could reduce the risk of DNA loss due to adsorption to surfaces. The method was validated using orthogonal measurement of stock template with quantitative HPLC, followed by quantification with ddPCR. The researchers also validated the stability of reference material plates and the homogeneity of randomly chosen plates.
Importantly, the NMI researchers also performed an inter-lab evaluation in which seven other labs were sent three sets of blinded samples of the reference material. Only one lab failed to achieve the desired level of precision and accuracy.
Significantly, the lab producing the failing result had an inexperienced user conduct the inter-lab experiment, Karlin-Neumann noted. "Thus, the testing exercise did indeed identify a lab where either instrumentation, reagent, or operator deficiencies did not allow high quality results to be achieved with the QX100/200 ddPCR system," he said. This highlights the need for such reference material in a clinical setting, as well as its value for facilitating more reproducible research results in the scientific community, he said. Furthermore, Karlin-Neumann pointed out that the better performing labs used the Bio-Rad C1000 thermocycler, "known for its highly uniform thermal block."
Reference materials for quality control are thought to be particularly essential in the so-called liquid biopsy space. As previously reported, a broad discussion of reference materials is ongoing in this field, and a number of new products have been recently launched to "test the tests" that aim to detect low levels of circulating tumor DNA.
Indeed, a study published last year proposing a QC method for dPCR using a different reference material noted that the most challenging part of the validation of dPCR methods is the verification of the "trueness," an expression of how close measured values come to a reference standard, as "representative samples with a reference value are often difficult to find."
The NMI reference material could be useful for clinical applications but could also help to improve reliability of other types of testing, for example genetically modified organism and environmental testing.
